Currently, most hierarchies are constructed based on document contents, which can be computationally challenging and memory intensive. More importantly, not all contents can be acquired due to memory capacity or privacy ...